I think the card's functionality depended on the vram installed in the machine. It couldn't, for instance, operate a maximum capacity in the 660av, because its vram isn't expandable.
Now that is interesting! I wonder how Apple lamed the 840AV's little cousin? If the circuitry and RAMDAC are nearly the same as in the 840AV you could be in business.
1) stacking VRAM ICs onto the 1MB of (soldered? ) VRAM ICs could work
2) you'd almost certainly need to patch the additional address line
3) if the RAMDAC is the same part you might be good to go?
4) if the RAMDAC is a less capable part
---- it could be similar to the Radius PrecisionColor 8-XJ vs. 24AC jailbreak attempt
______ part number/capability may be different, but they're likely to be pin compatible
______ might be necessary to double the data bus width in the case of the 8-XJ? Doesn't sound right somehow.
______ dunno about how that might go on the 660AV board?
___ SMT rework might fix the RAMDAC limitation of the 660AV
______ there seem to be plenty of DOA 840AV donor boards in the wild, I know mine is almost certainly in that category. :-/
@trag: Radius' PCI Thunders(?) were a simple rehash of the NuBus Card, that was the accusation held against them. Methinks there a breastplate from the armory to deflect the poignard's killing thrust might be.
That TI RAMDAC might be identified on any number of PC VidCards of the era as well.
h34r: